The project will focus on developing new measures that best detect abnormal and/or significant changes in networks.
A number of network similarity (dissimilarity/distance) measures have been developed in the literature. This project investigates the relationship between those measures and particular networks (protocols, routing algorithms, etc.) Possible topics of interest: Which measures detect best abnormal/significant changes in which networks (classified with respect to protocols, size, traffic variations, user population, etc.), what to do with multi-layered situations (physical/logical/social networks), comparisons of networks, localization of change - discover the areas/users of network that contribute most to change/abnormalities in network behaviour, what to do with massive networks? Which measures are more appropriate? The need to worry about computing times for certain measures? Should the algorithms be tweaked (parallelized) to speed up the processes?
The opportunity ID for this research opportunity is 984